Validation of Wind Tunnel Test and Cfd Techniques for Retro-Propulsion (Retpro): Overview on a Project Within the Future Launchers Preparatory Programme (Flpp)

VALIDATION OF WIND TUNNEL TEST AND CFD TECHNIQUES FOR
RETRO-PROPULSION (RETPRO): OVERVIEW ON A PROJECT WITHIN THE
FUTURE LAUNCHERS PREPARATORY PROGRAMME (FLPP)

D. Kirchheck, A. Marwege, J. Klevanski, J. Riehmer, A. Gu¨lhan

German Aerospace Center (DLR)
Supersonic and Hypersonic Technologies Department
Cologne, Germany

S. Karl
O. Gloth

German Aerospace Center (DLR) enGits GmbH
Spacecraft Department
Todtnau, Germany
Go¨ttingen, Germany

ABSTRACT

and landing (VTVL) spacecraft, assisted by retro-propulsion. Up to now, in Europe, knowledge and expertise in that field, though constantly growing, is still limited. Systematic studies were conducted to compare concepts for possible future European launchers [2, 3], and activities on detailed investigations of system components of VTVL re-usable launch vehicles (RLV) recently started in the RETALT project [4, 5].
Nevertheless, validated knowledge on the aerodynamic and aerothermal characteristics of such vehicles is still limited to a small amount of experimental and numerical investigations mostly on lower altitude VTVL trajectories, e. g. within the CALLISTO project [6, 7, 8]. Other studies were conducted to analyze the aerothermodynamics of a simplified generic Falcon 9 geometry during its re-entry and landing burns, but up to now without comparable experimental data [9, 10]. Consequently, numerical and experimental tools are not yet validated completely for the full range of a commercially relevant launcher trajectory, including high altitude control and propulsive maneuvers, as well as advanced aerodynamic control surfaces, like grid fins for example.

The project RETPRO, Validation of Wind Tunnel Test and
CFD Techniques for Retro-propulsion, aims at closing this gap

by preparing the necessary tools for a reliable design and simulation of a complete ascent and descent trajectory of an inline proposed launcher conﬁguration, which shall be based on current operational vehicles. For that purpose, the consortium, consisting of the Supersonic and Hypersonic Technologies Department and the Spacecraft Department of the DLR Institute for Aerodynamics and Flow Technology, and enGits, a supporting engineering company, providing efﬁcient computing resources and methods, follows a successive road-map:
The RETPRO project is a 2-years activity, led by the German Aerospace Center (DLR) in the frame of ESA’s Future Launchers Preparatory Program (FLPP), to close the gap of knowledge on aerodynamics and aero-thermodynamics of retro-propulsion assisted landings for future concepts in Europe. The paper gives an overview on the goals, strategy, and current status of the project, aiming for the validation of innovative WTT and CFD tools for retro-propulsion applications.

Index Terms— RETPRO, retro-propulsion, launcher aero-thermodynamics, wind tunnel testing, CFD validation

1. INTRODUCTION

The Future Launchers Preparatory Programme (FLPP) aims

at preparing the necessary components for future launch systems to be able to serve Europe’s needs for a secured and autonomous capability of access to space in the future. Primary goal is to reduce costs as well as improve ﬂexibility and availability to strengthen European launch operations and industry on the global market. Part of this endeavor, as stated by the European Space Strategy [1], is to supply competence and technology to allow the re-usability of future and existing spacecraft or components.
The paper at hand gives an overview on a 2-years activity, funded by the European Space Agency (ESA), under the framework of FLPP, dealing in particular with the making available of necessary tools for a detailed aerodynamic and aerothermal design of future European vertical take-off

The RETPRO project, under which the disclosed research and authorship was conducted, received funding by the European Space Agency (ESA) within the Future Launchers Preparatory Program (FLPP).

1. Deﬁnition of a reference conﬁguration, based on an operational launch system, performing VTVL with retro-

Flight / WTT

CFD

WP3: Computational Fluid Dynamics

CFD for Model Design
Numerical Rebuilding of WTT and Code-to-code Validation
Extrapol. to Flight

B
A

Numerical rebuilding / validation

High fidelity



1

WP1: Reference Case

and Test Conditions

WP2: Wind Tunnel Tests

Preparation of the Technical Work Packages
Wind Tunnel Model Design
Aerodynamic & Aerothermal Wind Tunnel Experiments

2
3

WP4: Joint Assessment of WTT/CFD and Flight Dynamics Simulation

Flight Simulation,

Mission Analysis

Continuous Updates of AEDB/ATDB

4
High fidelity



High fidelity / fitted inviscid CFD

D ^AEDB

WP5: Management Activities

C

Fig. 1. Work logic scheme

Fig. 2. Strategy for AEDB/ATDB generation

propulsion assisted descent and landing maneuvers.

of an AEDB/ATDB for ﬂight dynamics simulations. In indus-

try, those tools are often low-fidelity or simplified computational methods, derived from highly resolved Navier-Stokes computations. Usually, those simplifications are made in order to maintain efficient design processes and short response time. Obviously, the simplification of sophisticated computational methods needs to be validated, but also the use of high-fidelity CFD codes in new fields of application needs to be verified.
2. Generation of high-quality aerodynamic and aerothermal test data from wind tunnel experiments throughout the descent trajectory in the hypersonic, supersonic, and subsonic flight regimes with and without propulsive jet.

3. Numerical rebuilding of wind tunnel experiments for validation of the numerical models in order to allow a proper prediction of aerodynamic and aerothermal loads.
Therefore, a strategy of joint assessment of WTT and CFD is developed (Fig. 2). In Step 3 of the before mentioned road map, two high-fidelity CFD codes, the DLR TAU code and the DrNUM code by enGits GmbH (Sec. 6), are used to compare with each other, as well as WTT results, representing all relevant flight phases of a typical retro-propulsion assisted vertical descent and landing trajectory. Prior to that, in Step 2, The different flight phases are experimentally simulated in three complementary WTT facilities at DLR Cologne (Sec. 5). They are suitable for hypersonic, supersonic, and subsonic test conditions with additional cold and hot plume simulation capabilities using pressurized air or gaseous hydrogen and oxygen (GH2/GO2) combustion in combination with the external ambient flow. Throughout the various flight phases, the DLR TAU code is mainly used for rebuilding of WTT cases with hot plume simulation due to its capabilities in modeling chemically reacting gas mixtures and serves as reference for the calculations with the DrNUM code. The DrNUM code uses simplified hot gas modeling by an ideal gas mixture approach and offers highly parallellized computing capabilities, using graphics processing units (GPU)s, which represents a possible setup for an industrialized design process as mentioned above.
4. Numerical extrapolation to full scale flight conditions in order to generate each, a sophisticated aerodynamic

database and aerothermal database (AEDB/ATDB).

5. Simulation of the reference trajectories, enabling mission performance analysis and ﬁnal validation of the overall modeling approach.

After summarizing the project strategy, the paper elaborates on relevant currently operational launch systems and thereon based derivatives for reference configurations and mission trajectories. The definition of a concept for the experimental tests, numerical computations, and their joint assessment for a reliable extrapolation to full scale flight conditions is presented in line with the necessary experimental and numerical tools for the conduction of wind tunnel tests (WTT), CFD computations, and flight dynamics simulations. The internal work logic scheme, depicting the described road-map is shown in Fig. 1.

Based on a successful validation, both CFD codes are used in Step 4 to extrapolate the WTT results to a realistic flight configuration by performing full scale computations. The results of the full scale computations are gathered in order to sufficiently populate an AEDB/ATDB for a final 3-DoF or 6-DoF flight simulation, which is performed using the DLR

2. PROJECT STRATEGY

The main goal of this project focuses on the validation of tools, necessary for future design considerations in the field of supersonic and subsonic retro-propulsion and the generation in-house tool FDS (Flight Dynamics Simulator) (Sec. 7) in Step 5.
Since the first step of the project’s road map is to define reference configurations and missions based on operational launch systems, the derived flight simulations can finally be compared to original flight data of the reference cases in order to close the validation cycle.

3. SELECTION OF THE REFERENCE
LAUNCHER AND MISSIONS

(a)

(b)

(c)

(d)

3.1. State of the Art

There are a couple of concepts for reusable launch vehicle (RLV), existing in the United States, that have already been flown or tested to different degrees. Hence, they involve different technology readiness levels (TRL), as well as different ranges of mission envelopes and are therefore more or less sufficient as the project’s reference configuration. The Delta

Clipper Experimental (DC-X) vehicle (later NASA DC-XA),

a conceptual single-stage-to-orbit (SSTO) vehicle was tested in the 1990s up to an altitude of 3.1 kilometer (Fig. 3, a). Blue Origin’s New Shepard, a sub-orbital rocket for technology demonstration and space tourism, which reached an altitude of slightly above 100 kilometer, is sporadically operated since 2015 (Fig. 3, b). But only SpaceX’s Falcon rockets, i. e. the Falcon 9 and Falcon Heavy, are currently in regular commercial operation since 2013 and 2019 respectively (Fig. 3, c/d). Recently also SpaceX’s Starhopper, a technology demonstrator for new engine technologies to be used for future heavy lift rocket concepts is being tested since the middle of 2019. In addition to that, various concepts are being developed that have not been tested until now, but will use similar technologies. Blue Origin develops the New Glenn as a future heavy lift rocket, as also SpaceX develops the Falcon Super Heavy booster and it’s upper stage Starship.
In Europe, the implementation of such concepts started recently with the beginning of the HOMER (Airbus Defence and Space, ADS) [11, 12], EAGLE (DLR) [13] and FROG (CNES) [14] projects, representing small scale demonstra-

tors for Guidance, Navigation and Control (GNC) systems.

This is continued by the medium size launcher demonstrator project CALLISTO, which is a tri-lateral cooperation between CNES, DLR, and JAXA [6, 7, 8]. On European side, those concepts still implicate comparably low TRL for key technologies and a low to medium altitude range. Consequently, for the studies conducted in RETPRO, currently, only the Falcon 9 and Falcon Heavy booster return ﬂights are relevant for a sufﬁcient comparison in a broad MACH number and altitude range.
Fig. 3. State of the art in vertical take-off and landing reusable launch vehicle (VTVL RLV) in the United States: (a) DC-X, 1993; (b) New Shepard, 2016; (c) Falcon 9, 2015; (d) Falcon Heavy, 2019

Mission

NROL-76 SES-11

Flight Date
33 43
Type
01/05/2017 RTLS LEO 11/10/2017 DRL GTO
Orbit Payload
2,800 kg 5,200 kg

Table 1. Falcon 9 reference missions orbit (GTO). The LEO missions involve flight profiles with a landing of the first stage booster after a return to launch site (RTLS) near the launch pad. The booster landing of the GTO missions is performed as a downrange landing (DRL)

on an autonomous spaceport drone ship (ASDS) at a certain

downrange distance.
Obviously, the two approaches incorporate different trajectories and mission events for the descent flight phase and consequently different challenges to face for the flight simulation. The DRL, is typically performed, when more kinetic energy is needed for the orbit insertion, which is the case for GTO missions, or if larger payloads require the use of more fuel. In this case, not only the mission events, but also the flight profile differs considerably between RTLS and DRL missions. To study all parts of the different landing approaches, two reference missions are selected, representing both, an RTLS landing and a DRL.

The current candidates for the reference missions are the
NROL-76 mission as the reference for RTLS, and the SES- 11 mission for the DRL approach (Tab. 1). Trajectory reference data is taken from the SpaceX launch webcasts, which, in these cases, is provided for the first stage ascent and descent phases as altitude and total velocity over time (Fig. 4). For the SES-11 mission, the webcast is interrupted at approximately 430 seconds, making the final landing approach inaccessible. NROL-76 is therefore taken as the first reference for the validation of the launcher model.

3.2. Reference Missions

The Falcon 9 rocket is typically used by SpaceX for satellite

missions to low earth orbit (LEO) or geostationary transfer

200 150 100
50

200 150 100 50

Property Height / m Diameter / m
Stage 1
47.0 3.66
0.964
16
Stage 2
23.0
3.66 / 5.20
3.10

NROL-76 SES-11
RTLS

Nozzle exit diameter / m Nozzle expansion area ratio Empty mass / t Number of engines Oxidizer/propellant Speciﬁc impulse (sea level) / s Speciﬁc impulse (vacuum) / s Thrust (sea level) / kN Thrust (vacuum) / kN
165 4.4
1

DRL

26.55
9

LOX/RP-1

283.3 310.6
9×758.7 9×831.8

LOX/RP-1

60.0
342.0 160.0 915.0

0
-100

0

100

300

500

700

Downrange distance / km

Table 3. Properties of the reference launcher conﬁguration for the preliminary 3-DoF simulation
Fig. 4. Flight proﬁles of the candidate reference missions NROL-76 and SES-11, taken from the SpaceX launch webcast

Mission NROL-76 Falcon 9 v1.2 Block 3 B1032.1 Merlin-1D v1.2 SES-11 Falcon 9 v1.2 Block 3 B1031.2 Merlin-1D v1.2

Launcher

Booster

Booster Engine

Fig. 5. Shape of the generic reference launcher conﬁguration
Table 2. Reference launcher for NROL-76 and SES-11

sufficient for the definition of the experimental test matrix and test conditions and roughly indicates a proper definition of the reference launcher. The simulation will be updated in the course of the project as detailed aerodynamic data becomes available. This might also have an impact on the definition of the input parameters for the simulation, hence change the launcher definition.

3.3. Reference Launcher

For both reference mission candiates, the same Falcon 9 version, v1.2 or Full Thrust (FT), was used. Although, between Flights 33 and 39 a transition from the Block 3 to the Block 4 engine was made, the SES-11 mission was performed, using a previously ﬂown booster (B1031.2), which was refurbished after successful landing at the launch site in Flight 30. This was shortly before the NROL-76 launch, using booster B1032.1 (Table 2). This allows to use similar launcher models with simple adaptations regarding the payload and propellant masses for both missions and thus enables comprehensive validation possibilities.

5. WIND TUNNEL TESTING
5.1. Experimental Test Concept

For the design of the wind tunnel test concept, an experimental test trajectory is derived from the reference mission trajectory, taking into account the operating envelopes of the available test facilities. The goal is to represent a complete descent trajectory within the hypersonic, supersonic and subsonic flight regimes. Cold gas and hot gas plume simulation methods are additionally applied for the different retroburn maneuvers. An exemplary experimental test trajectory is shown in Fig. 8.
The exemplary descent trajectory comprises four major flight phases after entering the atmosphere. These are an aerodynamic descent in the upper atmosphere, followed by a reentry burn in order to limit the aerothermal loads at hypersonic to supersonic MACH numbers, a second aerodynamic descent phase in the supersonic regime, and the final landing burn, beginning approximately 30 s before touchdown and slowing down the vehicle from low supersonic through transonic and subsonic velocities. The aerodynamic descent in the lower atmosphere and the two retro-burn maneuvers are

4. DEFINITION OF THE GENERIC REFERENCE CONFIGURATION

From the selected reference missions and launcher, a generic launcher configuration is derived, which is used as the main reference in the project. In order to obtain a consistent design of the reference configuration, it is developed through preliminary 3-DoF simulations of the presented reference mission candidates. The current definition is based on the main properties, which are used as input data for the simulation and which are listed in Tab. 3. The underlying geometry, which will later be used for the wind tunnel model design and CFD calculations is depicted in Fig. 5. It is a generic representation of the reference launcher, including grid fins, landing structures, and a simplified base geometry.
The result of the preliminary flight simulation is shown in Fig. 6 and 7, compared to the webcast telemetry data. It is

200 150 100
50

represented by three wind tunnel facilities at DLR Cologne,

namely the Hypersonic Wind Tunnel (H2K), the Transonic Wind Tunnel (TMK) and the Vertical Test Section (VMK).

Throughout the whole test trajectory, various parameter variations are applied. Therefore, the different wind tunnel models are designed highly modular to enable variations of the type and deﬂection angles of control surfaces, thrust vector control angles, engine supply pressure, and nozzle expansion area ratio. Those can be combined with the application of continuous angle of attack polars in the H2K and TMK facilities, as well as discrete angle of attack variation in the VMK facility to create a broad parameter space.

ascent (3DoF) descent (3DoF) webcast

0

1

2

3

Velocity / km·s^-1

5.2. Hypersonic Re-entry Burn

Fig. 6. Preliminary 3-DoF simulation of the LEO reference
The H2K is used to duplicate the hypersonic re-entry burn ma-

neuver at discrete MACH numbers between 8.7, 7.0, 6.0, and 5.3. It is a blow-down type wind tunnel with a circular 600 millimeter free jet in an evacuated test section. In the H2K, the plume gas is simulated by high-pressure air supply, which is fed into the nose of the rocket model through a conventional sting support. A strain gauge with a tubular cross-section enables force measurements in four components, also in case of operation with plume simulation. Pressure and temperature are measured at the base of the rocket, along the main body and inside the pressure chamber. mission (NROL-76) compared to the SpaceX webcast data

200

ascent (3DoF) descent (3DoF) webcast

150 100
50
0

5.3. Supersonic Aerodynamic Descent

0

1

2

3

The TMK, a closed blow-down type wind tunnel, is used to duplicate the supersonic aerodynamic descent in the lower atmosphere between MACH numbers of 4.5 to transonic speeds by a continuous contour mechanism for the diverging nozzle part. The low supersonic range is limited due to the possibility of wave reﬂexions from the channel walls of the rectangular 600×600 millimeter test section, impinging on the wind tunnel models surface. In TMK, the same model can be used as for the H2K tests. Since, plume simulation is not necessary, it can be equipped with a common 6-component strain gauge, instead of the 4-component balance used in H2K.

Velocity / km·s^-1

Fig. 7. Preliminary 3-DoF simulation of the GTO reference mission (SES-11) compared to the SpaceX webcast data

200 150 100
50
10 8

re-entry burn aerodynamic descent aerodynamic descent

Ma_∞= 7.0
6.0 5.3

landing burn

H2K

6

5.4. Subsonic Landing-burn

TMK

The VMK facility is used to duplicate the subsonic part of the ﬁnal landing burn between MACH numbers of 0.9 to 0.3. It is an atmospheric free-jet facility operable in the supersonic and subsonic range using different nozzle assemblies. Its subsonic nozzle has an exit diameter of 340 millimeter and is located in the middle of a large 4×4.5 meter test section.

4

3.6

Altitude

3.0
2.5

VMK

Mach number H2K

2

1.7

TMK

0.8
0.3

VMK

0

270

320

370

420

470

520

The test section is prepared to conduct hot gas experiments using e. g. GH2/GO2 combustion in a model-integrated combustion chamber. Hence, more realistic rocket exhaust conditions can be reached in combination with the external wind tunnel ﬂow. These experiments are used to validate the hot gas and chemistry models of the different ﬂow solvers.

Time / s

Fig. 8. Experimental test trajectory, based on the candidate GTO reference mission (SES-11) webcast telemetry data; dots: trajectory extrapolated due to interrupted telemetry data

6. COMPUTATIONAL FLUID DYNAMICS (CFD)
6.1. Flow Solvers

6.2.1. TAU Computations

The models which will be applied for the thermodynamics and chemistry of the flow species will depend on the considered computational cases. In cold flow situations (static temperatures less then about 1000 K), air will be modeled as a calorically perfect gas (gas constant and ratio of specific heats are constant and independent of temperature and pressure). This applies to all cases without nozzle exhaust jets and with free stream MACH numbers below 3.5. At larger MACH numbers, an equilibrium air model will be applied. The thermodynamic properties vary correctly with temperature and pressure. Here, air is considered as an infinitely fast reacting mixture of N2, O2, N, O and NO. In cases with chemically reacting exhaust jets, air will be modeled as a mixture of thermally perfect gases with 76% mass fraction of N2 and 24% mass fraction O2. The properties of the single species vary with temperature and include high temperature effects such as vibrational excitation of the molecules.
The general approach for modeling exhaust gas will be based on a non-reacting (frozen) mixture of thermally perfect gases. The assumption of a non-reacting exhaust allows to define an equivalent single exhaust species whose thermodynamic properties are identical to the considered real mixture and vary only with temperature. This is achieved by the application of suitable mixture rules on the detailed exhaust composition. The exhaust species will include: CO, CO2, H2O, H2, OH, O2, H, O for JP-1 fueled engines and H2O, H2, OH, H2O2, HO2, H, O2, O for hydrogen fueled engines. The exact exhaust gas composition at the exit of the thrust nozzle will be determined by simplified thermodynamic and chemical analysis of the rocket engine using e. g. the NASA-CEA software.
Within the RETPRO project, two flow solvers, TAU and DrNUM, are used to full capacity in a complementary fashion in order to benefit from their specific qualities.
The TAU-code, developed by the DLR Institute of Aerodynamics and Flow Technology, is a well-established and widely used general purpose tool for a broad range of aerodynamic and aero-thermodynamic problems [15]. The TAU code is a second order finite-volume solver for the Euler and Navier-Stokes equations which includes a comprehensive range of RANS-based or scale resolving turbulence models and extensions for chemically reacting high enthalpy flows. It uses unstructured computational grids to facilitate the analysis of complex geometries and is highly optimized for the application on massively parallel high performance computing systems. TAU has been successfully applied to a wide range of subsonic to hypersonic flow problems, both in scientific and industrial applications, including the analysis of re-usable launcher configurations.
The DrNUM code has been firstly introduced to the public in the year 2013 and is jointly being developed by enGits GmbH and numrax GmbH. It is high-performance CFD software which makes extensive use of GPUs for massively parallel computing. It allows to process flow simulations in the order of tens of million cells on classical single node desktop computers. In contrast to classical grid methodologies (e. g. structured or unstructured) dual resolution grids are used, which consist of a set of coarsely arranged sub-grids, each with a highly resolved computational grid. This approach leads to achieve an optimal combination of high numerical and algorithmic efficiency within the sub-grids, while keeping geometric flexibility at the same time.
For sub-scale wind tunnel cases, the model engines are likely to be operated at very low oxidizer to fuel ratios which leaves a significant amount of unburnt hydrogen in the exhaust gas. As a preliminary study on the modeling approach for chemical reactions in the experimental test environment compared to real flight conditions shows, this hydrogen will ignite in the hot stagnation region of the exhaust jets and the resulting exothermic post-combustion will significantly increase the flow temperature and, hence, the thermal surface loads, which can be seen in Fig. 9 and 10. Therefore, chemical reaction between the ambient air and the exhaust jet of the rocket engines will be considered for the rebuilding of WTT.
The nature of the two codes makes them well suited for specific tasks within the joint assessment strategy of the RET- PRO project.